Discussion

In vivo vs. in vitro studies
In vivo studies are the priority in a research programme related to epigenetics and reproductive toxicity, because in vivo studies are the means by which robust systems with known adverse outcomes can be developed, and the context for understanding is established. Nevertheless, it was agreed that in vitro studies will be needed in order to complement the findings of in vivo work because they:
May provide epigenetic markers that could be used to complement the in vivo studies;
can help address mechanistic questions around causality;
can be used in the longer term to develop simple test systems for regulatory purposes that accommodate the 3R perspective;
it is important to choose adequate cell models that are able to mimic epigenetic rearrangements occurring during development, e.g. differentiation models. In contrast, tumour cell lines should be avoided as they are already perturbed, both genetically and epigenetically.
Regulatory relevance
Too early to augment test guidelines:
There are examples where test guidelines (e.g. OECD TG 421 and TG 422, 2015) have been augmented on the basis of a thorough literature review and data analysis.
Some participants indicated it is important to include adequate apical endpoints in epigenetic studies. This would already be an important signal as inclusion of epigenetic markers without inclusion of relevant apical endpoints would make the hazard interpretation impossible. They suggested that tissues could be collected and stored now (this could be a non-compulsory option) to be used as a tissue resource for retrospective analysis once epigenetics science has progressed. Such an approach, they said, would make better use of studies already performed and reduce animal use. Retrospective analysis of data from TGs is an acceptable approach at the OECD and is being increasingly utilised as part of the preparation for updating TGs with additional relevant endpoints. However, other participants, including industry and scientists from the USA, agreed that current understanding of toxicant-induced epigenetic change is still too limited to be formally incorporated into current test guidelines as a default requirement, and that more research is needed to demonstrate that examining epigenetic endpoints provides value in a regulatory context. Earl Gray from the US Environmental Protection Agency said: “It is too early at this time to augment current test guidelines, which have been used extensively and their value is known. We do not know what the added value of epigenetics measurements is”.
More data on chemicals of concern:
The participants from EU chemicals regulation emphasised that currently, there is insufficient information on the apical endpoint for a lot of chemicals. Regulators have poor data on chemicals of concern for generational effects. They advised that studies to collect these data should be conducted now in order to have the results in the future to more effectively regulate chemicals of concern.
Intergenerational vs. Transgenerational studies
Discussions around transgenerational effects (those observed in generations that are not directly exposed to the initial signal or environment that triggered the change), concluded that these effects should not be the focus of study at this time. This is because publications reporting toxicant-induced transgenerational effects in mammals were poorly reproducible and therefore the human health significance of reported toxicant-induced transgenerational effects has not been established. The consensus of opinion was that rather than focusing on transgenerational effects, a more appropriate starting point would be to investigate and understand the potential consequences of environmentally-induced epigenetic effects within a single generation, or between parent and child (intergenerational), and to understand somatic effects (epigenetically-mediated cellular memory of prior toxicant exposure within an individual). Particularly since these, or similar mechanisms, would also likely mediate hypothetical transgenerational effects.
Uncertainty Analysis: It was recommended that a systematic literature review with weight of evidence and uncertainty analysis would be required to underpin inter- and intra-lab variability, before designing the definitive models.
Determine which epigenetic alterations represent adverse changes, adaptive changes or are ‘biological noise’. Discussion centred on the fact that at this stage, we are not able to distinguish and cannot predetermine adaptive effects from adverse effects, especially when the causal link between epigenetic measurements and apical endpoints is not established.
Establish causality as much as possible: Reproducible epigenetic endpoints must be identified and examined within a mechanism/mode of action/adverse outcome pathway framework (including, but not limited to gene expression and histology) to establish a robust, mechanistically viable association between an epigenetic change and an in vivo adverse outcome. These relationships must be examined across time and the dose-response continuum.
Study design: There is a real need for reproducible and high confidence study designs with interpretable results. Too many epigenetic studies to date have lacked causality determination, were underpowered, and have used only a single, high dose-level exposure. Multiple dose levels must be examined to determine the dose-response relationships between an epigenetic endpoint and other molecular or apical endpoints. Careful analysis of cell populations should be performed to identify a true epigenetic alteration rather than an ‘epigenetic effect’ being caused by a change in cell ratio within the sample. The annex of the OECD 2012 review concluded that epigenetics studies to date were not sufficiently informative and recommended the need for definitive studies to better inform regulatory developments.
What epigenetic endpoints to measure: Discussions and presentations revealed that the following epigenetic endpoints are relevant: DNA methylation (including DNA hydroxymethylation) and histone post-transcriptional modifications of genes and miRNAs.